Fig. 1
From: Evaluation of classification in single cell atac-seq data with machine learning methods
Schematic overview of single-cell ATAC-seq. In data preprocessing step, raw sequencing data in.fastq format for each cell will be aligned first and stored in.bam format. Then we will sort bam files and remove duplicate reads in each cell. Finally we will integrate all files together to a merged and sorted bam file. In order to construct a cell-bin matrix, we will first call a fix-sized bins (window) list, then based on this peak file and bam file to construct a cell-bin matrix stored in dcCMatrix class. Both experiment will output predicted cell types for each method. These predicted cell types will be evaluated with ground truth cell types at the end in several aspect, including F1 score, confusion matrix, etc